Article

pubs.acs.org/JAFC

Nature of Phenolic Compounds in Coee Melanoidins

Carina Coelho, Miguel Ribeiro, Ana C. S. Cruz, M. Rosario M. Domingues, Manuel A. Coimbra,
Mirko Bunzel,# and Fernando M. Nunes*,

CQ-VR, Chemistry Research Centre, Department of Chemistry, Universidade de Tras-os-Montes e Alto Douro, 5000-801 Vila Real,
Portugal

QOPNA, Department of Chemistry, Universidade de Aveiro, 3810-193 Aveiro, Portugal

#
Department of Food Chemistry and Phytochemistry, Karlsruhe Institute of Technology, KIT, Karlsruhe, Germany
S Supporting Information
*

ABSTRACT: Phenolic compounds are incorporated into coee melanoidins during roasting mainly in condensed form (4262
mmol/100 g) and also in ester-linked form (1.11.6 mmol/100 g), with incorporation levels depending on the green coee
chlorogenic acid content. The phenolic compounds are incorporated in dierent coee melanoidin populations, but mainly in
those soluble in 75% ethanol (82%), a signicant correlation between the amount of phenolic compounds and the amount of
protein and color characteristics of the dierent melanoidin populations being observed. The incorporation of phenolic
compounds into coee melanoidins is a signicant pathway of chlorogenic acid degradation during roasting, representing 23% of
the chlorogenic acids lost. These account for the nearly 26% of the material not accounted for by polysaccharides and proteins
present in coee melanodins. The cleavage mechanism and the eciency of alkaline fusion used to release condensed phenolics
from coee melanoidins suggest that the phenolic compounds can be linked to the polymeric material by aryl-ether, stilbene type,
and/or biphenyl linkages.
KEYWORDS: coee, melanoidins, chlorogenic acids, phenolic, roasting, infusions

INTRODUCTION

Chemicals. 5-Caeoylquinic acid, caeic acid, ferulic acid,

coumaric acid, verartric acid, 2-deoxyglucose, fucose, galacturonic
acid, glucuronic acid, and caeine were obtained from Sigma (USA).
3,4-Dihydroxybenzoic acid, hydroquinone, catechol, gallic acid,
benzoic acid, 2-hydroxybenzoic acid, 4-hydroxybenzoic acid, 2,3dihydroxybenzoic acid, 3,5-dihydroxybenzoic acid, glucose, mannose,
arabinose, and galactose were purchased from Merck (Darmstadt,
Germany). 8-O-4-Dehydrodiferulic acid, 8,5(noncyclic)-dehydrodiferulic acid, and 5,5-dehydrodiferulic acid were isolated as described
in Jilek and Bunzel.17
Chlorogenic Acids (CGA) Extraction and Purication. Green
coee beans (200 g) were frozen in liquid nitrogen, ground, and
consecutively extracted with petroleum ether (24 h) followed by
methanol (24 h) in a Soxhlet apparatus. The petroleum ether extract
was discarded, and the CGA were recovered from the methanol extract
by rotary evaporation under vacuum at 35 C, dissolution in water, and
freeze-drying. The extract was puried by Sephadex G-25 column
chromatography, using water as eluent.18 The eluted solution
containing the chlorogenic acids was then treated with neutral lead
acetate (5 mL, 0.5 M), and the initial gray precipitate formed was
discarded. CGA were precipitated by further addition of neutral lead
acetate until no precipitate was formed.19 The precipitate was
dissolved in 5% sulfuric acid and reprecipitated at pH 8; this
precipitate was suspended in water, the remaining lead was removed
with the 5% sulfuric acid solution, and the lead-free ltrate was freezedried. The composition and purity of the CGA puried from green
coee beans was determined by RP-HPLC as described below.

Melanoidins are still today one of the most enigmatic food

macromolecules present in several heat-processed foods such as
coee, bread, malt, and beef,1,2 with coee brew being one of
the main sources of melanoidins in the human diet.3
The exact structure of coee melanoidins is still unknown,
but recent structural studies allowed a partial elucidation of
some of their structural features and highlighted their structural
diversity.410 Melanoidins importance is not limited to their
color contribution to heat-processed foods; these molecules are
also important for their modulation of avor release,
antioxidant and metal chelating properties, and dietary ber
behavior.1113 The involvement of chlorogenic acids or their
degradation products in coee melanoidins can be due to the
formation of either ester linkages14,15 or condensed forms
(non-ester covalent linkages) not yet completely structurally
disclosed.6,7,9 The relative amount of these two incorporation
forms is also not known, nor is the relevance of phenolic
compounds for the dierent properties of the melanoidin
populations present in coee brews.
The purpose of this work was to prole the dierent forms of
phenolic compounds present in coee melanoidins to assess
their abundance in the dierent coee brew melanoidin
populations. The inuence of green coee bean chlorogenic
acid levels on the amount of phenolic compounds incorporated
into coee melanoidins and in melanoidin color was also
studied. For this, a modied in bean coee model16 enriched
with a typical green coee chlorogenic acid mixture was used.
2014 American Chemical Society

MATERIALS AND METHODS

Received:
Revised:
Accepted:
Published:
7843

March
July 3,
July 5,
July 5,

31, 2014
2014
2014
2014

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

Preparation of the Modied in Bean Green Coee Models.

Modied in bean models enriched with CGA were prepared using a
previously developed method16 with small modications. Brazil arabica
green coee beans were initially freeze-dried until constant weight (6%
weight loss). For the preparation of the chlorogenic acid-enriched
coee model (CGAcoee), to 100 g of the freeze-dried green coee
beans was added 200 mL of a CGA aqueous solution (1.5 g/200 mL),
and the mixture was shaken for 16 h at 4 C. After incorporation, the
coee beans were washed with distilled water for removal of the
surface-adhering solution and freeze-dried. As a control, 100 g of
freeze-dried coee beans was immersed in 200 mL of water under the
same conditions.
Roasting of Coee Bean Model Samples. Prior to roasting, all
samples (original freeze-dried coee, water control, and CGAcoee)
were adjusted to the same moisture of 6% by spraying enough water
and checking coee weight after 1 week. Then, the coee beans (40 g)
were roasted at 200 C during approximately 5 min to obtain an
organic roast loss of 8%.
Preparation of Coee Brews and Isolation of the High
Molecular Weight Material (HMWM). Roasted coee beans were
ground (250450 m), and infusions were prepared by extraction of
25 g of powdered coee with 800 mL of water at 80 C during 20 min.
After ltration and vacuum concentration to 200 mL, 50 mL of coee
infusion was freeze-dried to calculate the total soluble solids. The other
150 mL was extensively dialyzed at 4 C (six water renewals, MW
cuto 1214 kDa). After dialysis, the retentate was freeze-dried, giving
the HMWM.
Fractionation of the HMWM by Graded Precipitation in
Ethanol. For the fractionation of HMWM for each coee infusion, the
method previously developed20 was used with a modication;16 for
complete solubilization of the freeze-dried HMWM, a solution of 6 M
urea was used. The HMWM was then precipitated by the addition of
absolute ethanol. The material was recovered in the 50% ethanol
solution (Et50) and in the 75% ethanol solution (Et75). It was
resolubilized in water and dialyzed (six water renewals, MW cuto
1214 kDa). After dialysis, the retentate was freeze-dried. The
material soluble in the 75% ethanol (EtSn) fraction was recovered after
concentration, dialysis, and freeze-drying.
Chlorogenic Acids and Caeine Content of Green and
Roasted Coees. CGA were extracted overnight from green and
roasted ground coees (1 g) with 100 mL of a solution of methanol/
water (70:30 v/v) containing 0.5% Na2SO3 with constant stirring (125
rpm) in the dark. To the solution was added 1 mL of veratric acid as
internal standard (100 mg/mL, in methanol). After ltration of the
coee powder, the colloidal material present in the extracts was
precipitated by the addition of 1 mL of Carrez I and II solutions.21 The
extracts were analyzed by HPLC (Dionex, Ultimate 3000) by injection
of 25 L of the sample on a reversed-phase HPLC column (C18-ACE;
25 cm length, 0.45 cm internal diameter, and 5 m particle diameter).
Eluent A was a 5% formic acid aqueous solution, and eluent B was
methanol. The eluent program was as follows: 05 min, 5% eluent B;
545 min, 40% B; 4565 min, 70% B; 6575 min, 5% B. The column
temperature was set at 25 C, and the ow was 0.8 mL/min. The
eluent was continuously monitored from 200 to 600 nm with a
photodiode array detector (PDA-100, Dionex). The identication of
caeine and 5-caeoylquinic acid was made by comparison of their
retention times and UVvis spectra with those of pure standards
analyzed under the same conditions. The other caeoylquinic acids
(CQA) and cinnamoyl-1,5--quinolactones (CQL) were tentatively
identied by comparison of their retention times and UVvis spectra
with those of the literature.22 CGA were quantied by the internal
standard method using a 5-CQA calibration curve. Caeine was
quantied using the same methodology and a caeine calibration
curve.
Characterization of Coee Bean Samples and Fractions.
Green and roasted coee and coee models were characterized
regarding their sucrose, glucose, and fructose contents by highperformance anion exchange chromatography.23 Total sugars content
was determined by anion exchange chromatography after Saeman and
acid hydrolysis.20 Protein content was determined according to the

Dumas method (PRIMACS, carbonnitrogen/protein analyzer,

Skalar, The Netherlands) by multiplying the nitrogen content by
6.25. The melanoidin content of fractions was determined by the Kmix
at 405 nm as previously described.4
Color of Roasted Coee Powder. Roasted coee powder colors
obtained for the original, control, and in bean CGAcoee model
were directly measured with a Minolta chroma meter (model CR-400,
Minolta, Tokyo, Japan). The equipment was set up for illuminant D65
and 10 observer. The equipment was calibrated with a white standard
(L* = 97.71, a+* = 0.59, and b* = 2.31), and each sample was put on
a Petri dish and read in ve dierent locations. Numerical values of
chroma (C*), hue (h), and color dierence (E*) were calculated
according to the following formulas:

b*
h* = tan1
a*
C* = (a*2 + b*2 )1/2
E* = [(L*)2 + (a*)2 + (b*)2 ]1/2
Total Phenolic Groups Content. The total phenolic groups
content of the coee samples was determined with the Folin
Ciocalteu reagent.4 For all coee samples, a 0.33 or 1 mg/mL solution
was prepared. To 1 mL of the coee sample solution was added 0.5
mL of FolinCiocalteu reagent. After mixing of the sample, 1 mL of a
saturated Na2CO3 solution was added, and ltered demineralized
water was added to reach a total volume of 10 mL. After 1 h of
reaction, the absorbance of the sample was measured at 725 nm. 5CQA was used as reference.
Adsorbed Phenolic Compounds. For determination of coee
melanoidins adsorbed phenolic compounds, a 5 mg/mL solution of
coee melanoidins was analyzed by direct injection (100 L) on a RPHPLC according to the method described previously for CGA
determination. For determination of the eect of melanoidins in the
determination of free phenolic compounds, melanoidin solutions (5
mg/mL) were spiked with dierent concentrations of 5-CQA and
incubated overnight at room temperature. Each solution was analyzed
by RP-HPLC as described previously, and the areas were compared
with that of aqueous solutions of 5-CQA with the same
concentrations.
Alkaline Saponication. For determination of ester-linked
phenolic compounds, melanoidins were subjected to alkaline
saponication by using a procedure that prevented oxidation of
phenolic compounds.24 To 750 L of the melanoidin solution (12
mg/mL) was added 750 L of a 2 M NaOH solution containing 2%
(w/w) ascorbic acid and 20 mM ethylenediaminetetraacetic acid. After
incubation for 1 h at 30 C, the mixture was quenched to pH 1 with
330 L of 5 M HCl to precipitate most of the coee material,
preventing precipitation during further analysis. The mixture was
stored for 2 h at 4 C, the precipitate was removed by centrifugation,
and the supernatant was analyzed by reversed-phase HPLC.
Experiments were performed at least in duplicate. The identication
of caeic, ferulic, and 4-hydroxycinnamic acids was made by
comparison of their retention times and UVvis spectra with that of
pure standards analyzed under the same conditions. Caeic, ferulic,
and 4-hydroxycinnamic acids were quantied by external calibration.
Alkaline Fusion. In a nickel crucible were weighed 1 g of solid
NaOH and 100 mg of zinc dust, and, after fusion of the mixture at 350
C,6,25 5 mg of HMWM was added (or 5 mg of melanoidin fractions).
After 10 s, the nickel crucible was removed and rapidly cooled on ice.
The fusion cake was solubilized by adding 6 M HCl, and 200 L of
internal standard solution was added (3,4-dimethoxybenzoic acid, 1
mg/mL) and acidied to pH 12 with 6 M HCl. The acidic mixture
was extracted four times with 30 mL of diethyl ether. After evaporation
of the organic solvent, carefully and with protection, the residue was
derivatized and analyzed by GC-EI/MS. Samples were derivatized with
0.5 mL of pyridine and 0.5 mL of N,O-bis(trimethylsilyl)triuoroacetamide and heated at 70 C for 30 min. After cooling to
room temperature, the silyl derivatives of phenolic compounds were
7844

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

Table 1. Soluble Sugar, Caeine, and Protein Contents (Grams per 100 g, Dry Basis) of Original, Control, and in Bean Green
and Roasteda Coee Models
coee
green
original
control
CGAcoee
roasted
original
control
CGAcoee
a

glucose

fructose

sucrose

total

caeine

proteinb

0.018 0.001
0.018 0.001
0.017 0.001

0.60 0.00
0.59 0.00
0.58 0.00

7.45 1.32
5.78 1.10
5.08 0.98

8.07 1.32
6.39 1.10
5.69 0.99

1.37 0.01
1.12 0.02
1.07 0.05

14.0 0.9
14.1 0.7
14.0 1.1

0.018 0.001
0.018 0.001
0.017 0.001

0.008 0.002
0.019 0.001
0.032 0.025

0.64 0.06
0.99 0.02
3.07 0.07

0.66 0.07
1.07 0.05
3.16 0.02

1.42 0.15
1.23 0.12
1.11 0.07

9.5 0.7
12.3 0.9
11.8 0.5

Not corrected for weight loss during roasting. b(Ntotal Ncaffeine) 6.25.

Table 2. Chlorogenic Acid (CGA) Composition of Original, Control, and in Bean Green and Roasteda Coee Model (Grams
per 100 g, Dry Basis)
green
CGA

control

roasted
CGAcoee

original

3-CQA
5-CQA
4-CQA
total CQA

0.53
3.36
0.007
3.90

0.02
0.02
0.001
0.03

0.44
2.92
0.001
3.37

0.03
0.17
0.001
0.20

0.44
3.34
0.006
3.79

0.01
0.20
0.002
0.20

0.48
1.04
0.17
1.70

1-FQA
3-FQA
5-FQA
4-FQA
total FQA

0.01
0.70
0.05
0.003
0.77

0.00
0.05
0.01
0.005
0.06

0.01
0.60
0.04
0.003
0.65

0.00
0.07
0.00
0.004
0.07

0.01
0.63
0.04
0.003
0.68

0.00
0.05
0.00
0.004
0.06

0.01 0.00
0.76 0.02
nd
nd
0.77 0.02

3-CoQA
5-CoQA
4-CoQA
total CoCQA

nd
0.007 0.001
0.04 0.02
0.05 0.02

3-CQL
4-CQL
3-FQL
4-FQL
1-FQL
total FQL

original

nd
nd
nd
nd
nd
nd

nd
0.008 0.003
0.03 0.02
0.04 0.02
nd
nd
nd
nd
nd
nd

nd
0.006 0.002
0.03 0.02
0.03 0.02
nd
nd
nd
nd
nd
nd

0.04
0.10
0.03
0.16

nd
nd
nd
nd
0.04
0.09
0.01
0.02
0.04
0.08

control
0.43
1.12
0.20
1.75

0.02
0.07
0.02
0.11

0.01 0.00
0.69 0.04
nd
nd
0.69 0.04
nd
nd
nd
nd

0.02
0.03
0.00
0.01
0.01
0.02

0.04
0.09
0.01
0.02
0.06
0.09

CGAcoee
0.53
1.07
0.11
1.72

0.00
0.02
0.00
0.02

0.01 0.00
0.88 0.01
nd
nd
0.89 0.01
nd
nd
nd
nd

0.00
0.01
0.00
0.00
0.00
0.01

0.04
0.09
0.04
Nd
0.03
0.07

0.00
0.00
0.05
0.03
0.02

3,4-diCQA
3,5-diCQA
total diCQA

0.18 0.00
0.30 0.01
0.48 0.01

0.16 0.02
0.26 0.05
0.42 0.07

0.31 0.01
0.48 0.01
0.79 0.02

0.01 0.00
0.07 0.01
0.08 0.01

0.01 0.00
0.07 0.01
0.08 0.01

nd
0.06 0.02
0.06 0.02

total (CGA)

5.19 0.13a

4.48 0.09ab

5.43 0.20bc

2.53 0.19

2.52 0.16

2.67 0.00

Not corrected for weight loss during roasting. Identical letters in the same row are signicantly dierent.

analyzed by GC-MS (Agilent) and quantied by GC-FID

(PerkinElmer) using a DB-1 column (30 m length, 0.23 mm internal
diameter, and 0.2 m lm thickness) and injecting 1 L in splitless
mode (time of splitless, 0.75 min). Both the injector and the transfer
line were set at 250 C. The initial column temperature was 70 C,
held for 1 min, increasing at 5 C/min until 250 C. The MS source
temperature was set at 180 C, and the electron ionization energy was
set at 70 eV, with scans from m/z 40 to 600. Retention times and
electron impact ionization spectra of pure standards were used to
identify the following compounds: 3,4-dihydroxybenzoic acid, hydroquinone, catechol, gallic acid, benzoic acid, 2-hydroxybenzoic acid, 4hydroxybenzoic acid, 2,3-dihydroxybenzoic acid, and 3,5-dihydroxybenzoic acid. Other phenolic compounds were identied by
comparison of their mass spectra with those of the Wiley mass
spectra database (Wiley 8). Quantication was performed by the

internal standard method using 3,4-dimethoxybenzoic acid as the

internal standard.
Statistical Analysis. All chemical analyses of green coees, roasted
coees, soluble solids, HMWM, and fractions obtained by ethanol
fractionation of the HMWM were performed in duplicate unless
otherwise stated. Signicant dierences (p < 0.05) in chemical
composition were analyzed by one-way ANOVA using the software
Statistica 8.0 (StatSoft, Inc., Tulsa, OK, USA). A NewmanKeuls post
hoc test was performed for detecting signicantly dierent means (p <
0.05). For calculating the correlation between the phenolic
compounds determined by the FolinCiocalteau method and those
released by alkaline saponication, the nonparametric Spearman
correlation coecient () and the parametric Pearson correlation
coecient (r) were used.
7845

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

Table 3. Total Sugara Composition (Grams per 100 g, Dry Basis) of Original, Control, and in Bean Green and Roasted Coee
Modelsb
green
original
control
CGAcoee
ANOVA
roasted
original
control
CGAcoee
ANOVA
a

Ara

Man

1.71 0.04
1.86 0.81
2.12 0.44
0.758

22.7 0.2
26.3 7.9
30.8 2.2
0.359

1.62 0.20
1.99 0.27
1.92 0.05
0.276

18.0 2.5
20.3 2.2
20.6 0.4
0.444

Gal
12.7 0.2
13.2 2.2
14.7 0.7
0.412
7.61 1.01
8.86 1.18
9.37 0.07
0.276

Glc

Glcpolymeric

total

totalpolymeric

11.2 1.0
9.81 3.58
10.3 1.3
0.835

7.61
6.60
7.44

48.3 1.1
51.2 4.5
57.9 4.6
0.169

44.7
48.0
55.0

8.41 1.23
9.30 0.80
10.2 0.2
0.259

8.07
8.75
8.60

35.7 5.0
40.4 4.4
42.0 1.1
0.370

35.4
39.9
40.4

Values expressed in anhydrosugar residues. bNot corrected for weight loss during roasting.

RESULTS AND DISCUSSION

With the purpose of studying the eect of the green coee
CGA content, if any, in the incorporation of phenolic
compounds into coee melanoidins, a modied in bean
model coee was produced by enrichment of the freeze-dried
coee beans with a CGA mixture isolated and puried from
green coee beans. The procedure for production of the
modied in bean coee model was performed at 4 C to
minimize side reactions. This was a modication from our
previous work, in which the incorporation was performed at
room temperature.16
Chemical Characterization of Green and Roasted
Coee Samples and Model. Green Coee. As can be seen
in Table 1, the sucrose hydrolysis was signicantly lowered in
this new protocol when compared to the previous one, when a
signicant sucrose hydrolysis was observed.16 Original green
coee had a total soluble sugar content of 8.1%, mainly
composed of sucrose (92%) and low amounts of glucose and
fructose (Table 1), which is in accordance with the literature
values.26 For the control coee there was a decrease of 20% of
total soluble sugar content when compared with the original
green coee, only due to the decrease in sucrose. For the
CGAcoee model the amounts of sucrose, fructose, and glucose
are similar to those of the control coee (Table 1). This
decrease in the amount of soluble sugars can be due to leaching
of compounds during the incorporation process and contrasts
with the previous procedure at room temperature,16 when all of
the solution was incorporated into coee beans.
The amount of caeine present in the original coee was
signicantly higher than that present in control and CGAcoee
model, but the protein content of the coees, estimated on the
basis of the nitrogen content after correction for caeine
nitrogen, was not signicantly dierent (Table 1).
The amount of total CGA in the CGAcoee model was
signicantly higher (21% higher) than that present in control
coee, showing that the incorporation process was successful
(Table 2). As observed for the soluble sugars, control coee
contained a signicantly lower amount of CGA. The CGA
proles were similar in all coees, with 5-CQA as the most
abundant, representing, on average, 64% of the total CGA.
The sugar composition of green coee polysaccharides was
determined by subtracting from the total sugar content the
contribution of glucose and sucrose present as soluble sugars
(Table 3). The sugar content was in accordance with the
known sugar composition of coee cell wall polysaccharides,16,27,28 not signicantly dierent for the original, control,
and CGAcoee model.

The results presented showed that this new protocol for

production of the modied in bean model coees results in
fewer changes in the chemical composition of control coee in
relation to the original coee when compared with the original
protocol,16 being more suitable for the study of the
incorporation of selected compounds into green coee beans
prior to the roasting process.
Roasted Coee. Original, control, and CGAcoee model
coees were roasted to 8.0 0.5% of organic matter loss, and
their chemical composition was determined. After the roasting
process, losses of 92% of the total soluble sugar content of the
original coee, 84% for the control, and only 44% for the
CGAcoee model were observed (Table 1). This dierence in
total soluble sugar loss may be due to the higher content of
CGA in the CGAcoee model that might be preferentially lost
during the roasting process (further discussed below). The
decrease in the total soluble sugar in roasted original and
control coees are similar to that described in the literature.16,29
The caeine level remained unchanged, and the protein
content of roasted coees decreased between 20 and 38%
(Table 1).
The amount of total CGA also decreased during the roasting
process in all coee models (Table 2). For the original and
CGAcoee model, the amount of total CGA decreased 55%,
and a decrease of 48% was observed for control coee. This
decrease is in accordance with a coee submitted to a degree of
roast of 8%.30
The content of polymeric sugars had also a signicant
decrease after the roasting process (Table 3). All coee samples
behaved similarly, with an average decrease of 28% in the
content of total polymeric sugars, in accordance with the
literature.31,32 For all coees, when calculated for the equivalent
weight of green coee, the polymeric glucose levels were
maintained or decreased slightly. For all other sugars, when
calculated for the equivalent weight of green coee, there was
observed a decrease (17% for arabinose, 30% for mannose, and
41% for galactose).
The identical degradation prole of the studied coee
constituents led to the conclusion that the freeze-drying of
green coee beans and their immersion in water did not change
signicantly the behavior and properties of coee during the
roasting process.
During the roasting process, browning reactions occur,
contributing to the coee brown color formation. By visual
inspection of the interior of roasted coee beans, an even
distribution of the brown color could be observed. The
CIELAB color space used to characterize the coee powders
7846

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

7847

18.4
3.86
1.44
0.26
2.16

14.9
3.73

22.5
2.39

37.3
6.7
56.0

32.7
16.8
50.5

34.8
12.5
52.7

g/100 g HMWM

1.3b
0.1
3.8
1.8

0.3a
0.9
0.6
1.2

0.0121

5.94 0.6ab
1.70 0.4
11.3 1.7
13.2 0.9

11.5
2.7
19.2
16.7

11.1
3.9
14.2
16.8

Ara

0.234

20.1
38.7
11.1
3.5

17.1
34.9
7.9
4.0

20.2
30.1
11.6
5.5

Man

2.0
9.2
1.9
0.2

1.9
1.1
1.7
0.4

0.3
3.0
0.2
0.4

0.7a
2.0
2.6
1.2

1.3ab
0.8
3.7
0.3

0.0296

12.5
3.5
25.3
14.8

17.2 1.9b
4.85 0.22
30.4 5.5
20.4 2.2

19.8
9.4
32.8
23.8

Gal

0.15ab
1.27
0.08
0.05

0.09a
0.09
0.07
1.48

0.0277

ab
1.02 0.28
1.13 0.31
0.87 0.48

0.45
1.54
0.49
0.44

0.50
0.53
0.50
1.63

Glc

4.0
10.1
7.6
1.5

5.2
0.2
11.0
4.4

1.3
0.3
3.4
4.4

0.0940

38.6
44.9
48.7
32.3

46.2
44.0
57.9
41.6

51.6
44.0
59.1
47.7

total

1.0ab
0.4
0.5
0.7

0.8ab
0.2
0.5
0.3

0.3a
0.1
0.5
0.4

0.0030

18.9
4.8
14.7
27.9

14.1
5.2
12.3
14.6

9.9
2.3
12.1
14.3

protein

Phenolics measured by the FolinCiocalteu method and expressed as 5-CQA equivalents. Identical letters in the same column are signicantly dierent.

ANOVA

original
soluble solids
HMWM
Et50
Et75
EtSn
control
soluble solids
HMWM
Et50
Et75
EtSn
CGAcoee
soluble solids
HMWM
Et50
Et75
EtSn

g/100 g coee
0.03
0.05a
0.01
0.02
0.11
0.08
0.00ab
0.03
0.02
0.19
0.00
0.15b
0.09
0.13
0.07

0.0234

0.70
1.27
0.56
0.88
1.27

0.74
0.89
0.89
0.56
1.07

0.75
1.40
0.85
0.82
1.37

Kmix405 nm

2.99
1.11
2.40
3.19

2.24
1.75
1.88
2.44

3.63
1.58
2.85
3.61

MBI

0.049

6.19
1.55
3.13
8.20

4.50
1.57
1.93
7.28

4.59
1.16
1.69
5.36

0.36ab
0.16
0.27
0.45

0.61b
0.04
0.13
0.18

0.24a
0.10
0.20
0.33

phenolicsa

Table 4. Yield and Chemical Composition (Grams per 100 g, Dry Basis) of the High Molecular Weight and Resulting Ethanol-Precipitated Fractions of the Original, Control,
and in Bean Coee Models

Journal of Agricultural and Food Chemistry

Article

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

showed that the C* value (chroma) of the CGAcoee model

was signicantly higher than that of the control, showing that
the incorporation of CGA increased the color intensity of the
resulting roasted coee. Furthermore, this increase is related to
a signicant increase of the b* value for the CGAcoee model,
showing that the CGAcoee model had a yellower color than
that presented by the control.
Chemical Composition of Coee Infusions and
HMWM. Coee infusions were prepared from the original,
control, and CGAcoee model, and the HMWM was isolated
from each coee infusion. The amount of soluble solids
extracted from control coee was lower (33% less) than that
extracted for the original coee (Table 4). For the CGAcoee
model an increase of 23% was observed in the amount of
soluble solids in relation to control coee. However, using the
Kmix 405 values for estimating the melanoidins present in soluble
solids, it can be observed that the incorporation of CGA did not
signicantly increase the Kmix 405 value of the soluble solids in
relation to the control.
The chemical composition of the HMWM isolated from
coee infusions prepared from the original, control, and
CGAcoee model is shown in Table 4. The main component
of all the HMWM was the carbohydrates (3952%), with no
signicant dierences between the coees, followed by the
protein (1019%), with CGAcoee model containing a higher
amount of protein when compared to the control (+34%). The
amount of phenolic compounds was also signicantly higher for
the CGAcoee model when compared to the control (+38%),
with the latter presenting a similar amount of phenolic
compounds to that present in the original coee HMWM.
From these results it is evident that the incorporation of CGA
in the in bean model resulted in an increase in the amount of
phenolic compounds and protein in the HMWM.
The original coee HMWM presented a signicantly higher
Kmix 405 nm value when compared to control coee that also
presented a signicantly lower Kmix 405 nm value than the
CGAcoee model. In fact, the proportion of Kmix 405 nm,
indicative of the brown color of the HMWM, in relation to
the Kmix 405 nm of the soluble solids was higher for the
CGAcoee model (37%), being the lowest for the control
coee (19%). By subtracting the percentage of carbohydrates
and protein present in the HMWM, the amount of unknown
material can be calculated. The results show that it is similar for
all coees and has an average value of 60%. Also, and as
previously observed,16 there is no trend between the amount of
unknown material and the color of the HMWM determined by
the Kmix 405 nm value. Using the melanoidin browning index
(MBI), calculated by dividing the Kmix 405 nm values by the
relative amount of unknown material,16 which corresponds to
the absorbance at 405 nm by unit amount of unknown material
(Table 4), it can be seen that the enrichment of CGAcoee
with CGA increased by 33% the MBI. These observations allow
inferring that the CGA have a signicant inuence on the color
of the HMWM, although not increasing signicantly the
amount of unknown material.
Proling Phenolic Compounds in Coee HMWM. To
evaluate the dierent natures of phenolic compounds present in
the HMWM, either adsorbed, ester-linked, or in condensed
form, as well the eect of the green coee CGA levels in the
phenolic compounds prole, the HMWM was subjected to
alkaline saponication and alkaline fusion.
Adsorbed Phenolic Compounds. The amount of adsorbed
phenolic compounds was determined by direct injection of the

HMWM solutions on an RP C-18 column. To determine the

possible inuence of the HMWM matrix on the determination
of phenolic compounds, a solution containing 5 mg/L HMWM
and dierent amounts of 5-CQA was compared to the amount
determined by injection of the same amounts of 5-CQA in
water. The regression line obtained by representing the amount
determined in water in relation to the amount determined in
the presence of HMWM was y = 0.9962x + 0.2584 with r =
0.9998, indicating that the presence of HMWM did not
interfere in the determination of phenolic compounds by RPHPLC. The direct injection of HMWM showed the presence of
intact CGA acids in all HMWMs (Figure 1a), mainly 5-CQA,
although the presence of diCQA and 3CQA could also be
detected (not shown). The amount of adsorbed CGA ranged
from 0.1 to 0.7 mmol/100 g HMWM.

Figure 1. Prole of phenolic compounds present in (a) original,

control, and CGAcoee model HMWM; (b) original, control, and
CGAcoee model 75% ethanol soluble melanoidin population.

Ester-Linked Phenolic Compounds. The presence of esterlinked phenolic compounds in the HMWM from original,
control, and CGAcoee was determined after saponication.
This procedure, known to release ester-linked phenolic
compounds, 33 allowed the release of simple phenolic
compounds from the HMWM, including caeic, ferulic, and
coumaric acids (Figure 2). Caeic acid was the main cinnamic
acid detected in all analyzed HMWMs. These simple phenolic
compounds released by alkaline saponication have their origin
in the adsorbed CGA present but also, probably, in the CGA
incorporated into melanoidins, with structures where the
cinnamic acids are esteried and maintain unchanged their
structure. The amount of phenolic compounds released by
saponication ranged from 1.1 to 1.6 mmol/100 g, which was
signicantly higher than that found adsorbed in the HMWM.
For calculating the amount of ester-linked phenolic compounds
in the HMWM, the amount of simple phenolics released by
saponication was subtracted from the amount of adsorbed
phenolics (Figure 1a). As can be observed, the amount of esterlinked phenolic compounds present in the HMWM ranged
from 0.5 to 1.4 mmol/100 g. The values are in the range of
those found in a previous study,15 although lower values have
been reported14 or even not detected after alkaline hydrolysis,6,9 probably due to their oxidation during extraction.24 The
amount of ester-linked phenolic compounds present in the
CGAcoee model was signicantly higher than that present in
the HMWM of control, showing that the incorporation of
phenolic compounds in the CGAcoee resulted in a
signicantly higher amount of ester-linked phenolic compounds
in the HMWM.
7848

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

Figure 2. Phenolic compounds ester linked to the HMWM.

Figure 3. Nature and relative abundance of phenolic compounds released by alkaline fusion of the high molecular weight material recovered from
the in bean CGAcoee model.

Condensed Phenolic Compounds. For determining the

amount of condensed phenolic compounds present in the
HMWM, the samples were subjected to alkaline fusion. As can
be observed in Figure 3, more than 10 simple phenolic

compounds were released from the HMWM after alkaline

fusion. 3,4-Dihydroxybenzoic acid was the most abundant
phenolic compound in all samples. The abundance of the
dierent phenolic compounds, expressed as the relative areas,
7849

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

sition, including melanogalactomannans, melanoarabinogalactans, and melanoidins containing low amounts of carbohydrates.6,7 These macromolecules can be roughly separated by
ethanol precipitation. Galactomannans and melanogalactomannans can be obtained by precipitation in 50% ethanol
solutions (Et50), arabinogalactans and melanoarabinogalactans
can be obtained by precipitation in 75% ethanol solutions
(Et75), and a second melanoarabinogalactan fraction and the
melanoidins with low amounts of carbohydrates remain soluble
in 75% ethanol solutions (EtSn). The yield and chemical
composition of the fractions obtained by ethanol fractionation
are shown in Table 4. For all coees, the amount of material
recovered was highest on the EtSn fraction (53%, on average)
followed by the Et50 fraction (35%), and the Et75 fraction
accounted for the lowest amount (13%). This distribution of
the dierent melanoidin populations is in accordance with
previous ethanol fractionation procedures in Arabica and
Robusta coees.6,16,20,34,35 The amount of protein recovered
in the fractions isolated from the dierent coees increased
signicantly with the increase of the ethanol concentration, in
accordance with Bekedam et al.4 The amount of protein
recovered in the EtSn and Et75 fractions from CGAcoee was
signicantly higher than that of control. The phenolic
compounds present in the each fraction also increased with
the increase in ethanol solubility, being also observed a
signicantly higher amount of total phenolics in the Et75 and
EtSn fractions recovered from CGAcoee when compared to
control. A signicant correlation between the amount of
protein and phenolic compounds present in the ethanol
fractions (r = 0.830, p < 0.05) was also observed. When the
percentage of phenolic compounds present in each fraction is
calculated in relation to the amount of phenolic compounds
present in the HMWM, it can be observed that the phenolic
compounds present in the EtSn fraction represent nearly 83%
of the HMWM phenolic compounds, followed by the Et50
fraction (15%) and the Et75 (12%). These results show that the
melanoidin population soluble in ethanol contained the
majority of the phenolic compounds incorporated into the
HMWM during roasting, which is in accordance with Nunes et
al.16 and Bekedam et al.4 The MBI increased for all coee
fractions with the increase of their solubility in ethanol. For the
Et75 and EtSn fractions recovered from the CGAcoee model,
an average of 30% increase in the MBI was observed. In these
fractions, the amounts of protein and phenolic compounds are
signicantly higher for CGAcoee model when compared to
the control. Also, a signicant correlation between the phenolic
compounds and MBI (r = 0.674, p < 0.05) is observed. The
amount of phenolic compounds was also signicantly
correlated with the color of the dierent melanoidin
populations (r = 0.782, p < 0.05), a signicant correlation
between the amount of protein and MBI (r = 0.749, p < 0.05)
being also observed.
As the melanoidin populations soluble in 75% ethanol
contained the majority of the phenolic compounds present in
the HMWM, the nature of the phenolic compounds present in
these melanoidin populations was further studied (Figure 1b).
The EtSn fraction did not contain adsorbed phenolic
compounds, which can be explained by the experimental
procedure used, including dialysis and urea solubilization,
probably resulting in the loss of the adsorbed phenolics. The
amount of ester-linked phenolic compounds was in the range of
that found for the HMWM, and there was no signicant
dierence between the amounts of ester-linked phenolics for

was not signicantly dierent for the dierent HMWMs

studied. The amount of phenolic compounds released by
alkaline fusion was signicantly higher than the amount of
phenolic compounds released by alkaline saponication by a
factor of approximately 40 times and ranged from 42 to 62
mmol/100 g. These values are in the range of those reported by
Nunes et al.16 and Nunes and Coimbra6 for HMWMs isolated
from coee infusions and higher than those found by Takanake
et al.9 for a zinc chelating fraction isolated from instant coee.
The low abundance of esteried phenolic compounds if
compared to condensed phenolics present in coee melanoidins is in accordance with the results of Gniechwitz et al.,10 who
performed an NMR analysis of an isolated intact melanoidin
fraction and could not observe HSQC signals diagnostic for
intact caeic acid moieties. When the amount of phenolic
compounds present in condensed form was calculated by
subtracting the amount of phenolic compounds released by
alkaline fusion from that released by alkaline saponication, it
can be observed that the amount of phenolic compounds
present in condensed form in the HMWM recovered from the
CGAcoee was signicantly higher than that present in control.
As observed for the ester-linked phenolic compounds, the
incorporation of phenolic compounds in CGAcoee model also
increased the amount of condensed phenolic structures in the
HMWM (Figure 1a). Expressing the amount of condensed
phenolic structures in equivalent weight of caeic acid, it can be
concluded that the HMWM contained 7.5 g of condensed
phenolics/100 g for control coee, 10.0 g/100 g for the original
coee, and 11.2 g/100 g for the CGAcoee model, showing
that these condensed phenolic structures are an important
component of the HMWM. In fact, the contribution of these
condensed phenolic structures to the overall weight of the
HMWM can be higher as some of the phenolic compounds
bound to the HMWM seem to remain esteried with quinic
acid after roasting.14
Mechanism and Eciency of Alkaline Fusion. The alkaline
fusion method is known to release phenolic compounds from
condensed structures such as lignin and humic acids.25
Nevertheless, there is a lack of knowledge about the nature
of linkages between phenolic compounds cleaved by alkaline
fusion and the eciency of their cleavage. To have a deeper
insight into which types of linkages are cleaved by alkaline
fusion and to what extent, a series of standard phenolic
compounds containing aryl-ether linkages (8-O-4-dehydrodiferulic acid), stilbene-type linkages (8,5(noncyclic)-dehydrodiferulic acid), and biphenyl linkages (5,5-dehydrodiferulic acid)
between ferulic acids were subjected to alkaline fusion. The
phenolic compounds containing aryl-ether linkages and
stilbene-type linkages were quantitatively cleaved by alkaline
fusion, yielding only 3,4-dihydroxybenzoic acid, catechol, and
hydroquinone. Nevertheless, the phenolic compound containing biphenyl type linkage was not quantitatively recovered from
the alkaline fusion, being recovered only 19% of the original
phenolic compound. This result may show that this type of
linkage is more resistant to alkaline fusion. Also, it is possible
that the phenolic compounds released were of dierent nature
or that they were lost during the alkaline fusion, not allowing
their recovery.
Fractionation of the HMWM by Ethanol Precipitation.
The HMWM of roasted coee infusions is composed by a
mixture of thermally transformed galactomannans and
arabinogalactans and by a diversity of melanoidin populations
presenting dierent physical properties and chemical compo7850

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

cleaved by alkaline fusion. The presence of phenolic

compounds besides 3,4-dihydroxybenzoic acid, catechol, and
hydroquinone in the alkaline fusion may represent dierent
linkages or phenolic compounds transformed during roasting
with structural features dierent from caeic acid, but not yet
available to be determined. The detection of 4-hydroxybenzoic
acid in the alkaline fusion medium may represent the
incorporation of coumaric acid in coee melanoidins. The
relative contribution of ferulic acid and caeic acid to the
amount of phenolic compounds incorporated into coee
melanoidins cannot be deduced due to the observed
demethylation of phenolic compounds during alkaline fusion.16
Further work is needed to disclose the structural details of the
condensed phenolic compounds present in coee melanoidins.
The correlation between proteins and phenolic compounds in
the dierent melanoidin populations suggests that the
incorporation of phenolic compounds in coee melanoidins is
related to the amount of proteins. This can be explained by the
easy cross-link between phenolic compounds through quinoidal
structures and amino acids such as lysine and cysteine present
in proteins.4042 The condensation of phenolic compounds
during roasting has been shown to occur, but only low
molecular weight structures have been found.4345 These might
represent the starting point of the reaction that, on further
reaction with proteins and also possibly with polysaccharides,
incorporates these compounds into coee melanoidins. As
condensed phenolic structures are present in light-roasted
coee,6 the incorporation process seems to occur as soon as the
roasting starts, and this is in agreement with the data obtained
from Smrke et al.46 The presence of phenolic compounds in
coee melanoidins, mainly in condensed form, may be an
explanation for the high antioxidant activity described for coee
melanoidins.47,48 The association of these condensed phenolic
structures with indigestible polysaccharides and melanoidins
present in coee HMWM may also be one of the reasons for
the observed association between coee consumption and the
antioxidant capacity of feces from healthy subjects.49 The
presence of these condensed phenolic structures can act as a
radical-sponge in the gastrointestinal tract (stomach and small
and large intestine), as those described for other antioxidant
dietary bers when associated phenolic compounds are
present.50 Studies on the impact of gut microorganisms on
the antioxidant activity of melanoidins are rare, but it was
demonstrated, for example, that after a 24 h in vitro
fermentation of high-molecular weight coee fractions with
gut microorganisms, roughly 25% of the original antioxidant
activity was still present.13 Due to the relatively high amount of
phenolic compounds that can reach the gut, especially the large
intestine, the fact that they are linked to the coee dietary ber
deserves further study.

the three coees studied. The amount of condensed phenolic

compounds present in the EtSn fraction was signicantly
higher, by a factor of 60, than the ester-linked phenolic
compounds. The amount of condensed phenolic compounds
present in the CGAcoee model was, on average, higher than
that present in the control coee (Figure 1b). Nevertheless, it
was not signicantly dierent, probably due to the higher
variability obtained in the analysis, as the total phenolic acids
measured by the FolinCiocalteu colorimetric method were
signicantly higher. There was observed a signicant correlation
between the amount of phenolic compounds released by
alkaline fusion and that determined by the colorimetric Folin
Ciocalteu method ( = 0.886, p < 0.05), although this
correlation was not signicantly linear (r = 0.767, p < 0.075).
The FolinCiocalteu colorimetric method measures the
samples reducing capacity, which includes, among others,
phenolic compounds.36 However, this assay is also sensitive to
the presence of proteins, ascorbic acid, and other reducing
compounds.37,38 In fact, melanoidins produced from model
systems containing only sugar and amino acids are known to
possess reducing capacity,39 showing that proteins or reducing
Maillard reaction products such as reductones present in the
material can contribute to the reducing power of the solution,
resulting in an overestimation of the amount of phenolic
compounds present when measured by the FolinCiocalteu
method. Nevertheless, as the reducing power of 5-CQA is 21
times higher, on a weight basis, than that of bovine serum
albumin (BSA), protein interference should be comparably low
(although the actual interference clearly depends on the amino
acid composition of the proteins). Also, it is certainly harder to
estimate the contribution of Maillard reaction products to the
reducing power of the actual coee sample. Nevertheless,
although the application of the FolinCiocaultau method can
have drawbacks, it is widely used for accessing the presence of
phenolic compounds in coee melanoidins, and the results
obtained in this work show that it can be used as an indicator
for phenolic compounds present.
Fate of CGA during Roasting, Contribution to
Melanoidin Formation, and Potential Health Eects.
The amount of phenolic compounds incorporated in coee
melanoidins, determined by the alkaline fusion method, ranged
from 10 to 13 mol/100 mol of the CGA content present in
green coee. When the amount of phenolic compounds
incorporated in coee melanoidins was calculated in relation
to the amount of phenolic compounds lost during roasting, the
values ranged from 18 to 33 mol/100 mol. From these values it
can be observed that the incorporation of phenolic compounds
into the HMWM is a signicant pathway of CGA degradation
during roasting. When the amount of phenolic compounds
incorporated in the HMWM is expressed as caeic acid, it
represents, on average, 23 g/100 g of the material not
accounted for by polysaccharides and proteins. These results
show that the phenolic compounds incorporated into coee
melanoidins represent a signicant proportion of the previously
described unknown material. The correlation found between
the amount of phenolic compounds present in the dierent
melanoidin populations and the Kmix color as well the MBI
shows that the incorporation of phenolic compounds has an
impact on the color of melanoidins. Although the structure of
phenolic compounds present in the coee melanoidins cannot
be deduced from the data obtained from the methods applied,
the data showed that they can be connected by aryl-ether,
stilbene-type linkages, and biphenyl linkages, as these can be

ASSOCIATED CONTENT

* Supporting Information
S

Additional gures and tables. This material is available free of

charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION

Corresponding Author

*(F.M.N.) E-mail: fnunes@utad.pt. Phone: +351259350242.

Funding

We appreciate the nancial support provided to the project

PTDC/QUI-QUI/100044/2008, to the Research Unit in Vila
7851

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

(21) Trugo, L. C.; Macrae, R. A study of the effect of roasting on the

chlorogenic acid composition of coffee using HPLC. Food Chem. 1984,
15, 219227.
(22) Farah, A.; de Paulis, T.; Moreira, D. P.; Trugo, L. C.; Martin, P.
R. Chlorogenic acids and lactones in regular and water decaffeinated
arabica coffees. J. Agric. Food Chem. 2006, 54, 374381.
(23) Abraao, A. S.; Lemos, A. M.; Vilela, A.; Sousa, J. M.; Nunes, F.
M. Influence of osmotic dehydration process parameters on the quality
of candied pumpkins. Food Bioprod. Process. 2013, 91, 481494.
(24) Nardini, M.; Cirillo, E.; Natella, F.; Mencarelli, D.; Comisso, A.;
Scaccini, C. Detection of bound phenolic acids: prevention by ascorbic
acid and ethylenediaminetetraacetic acid of degradation of phenolic
acids during alkaline hydrolysis. Food Chem. 2002, 79, 119124.
(25) Batistic, L.; Mayaudon, J. Etude chromatographic des acides
phenoliques et des phenols obtenus par la fusion alcaline des acides
humiques. Plant Soil 1970, 33, 473477.
(26) Murkovic, M.; Derler, K. Analysis of amino acids and
carbohydrates in green coffee. J. Biochem. Biophys. Methods 2006, 69,
2532.
(27) Fischer, M.; Reimann, S.; Trovato, V.; Redgwell, R. J.
Polysaccharides of green Arabica and Robusta coffee beans. Carbohydr.
Res. 2001, 330, 93101.
(28) Oosterveld, A.; Harmsen, J. S.; Voragen, A. G. J.; Schols, H. A.
Extraction and characterization of polysaccharides from green and
roasted Coffea arabica beans. Carbohydr. Polym. 2003, 52, 285296.
(29) Hughes, W. J.; Thorpe, T. M. Determination of organic acids
and sucrose in roasted coffee by capillary gas chromatography. J. Food
Sci. 1987, 52, 10781083.
(30) Perrone, D.; Donangelo, R.; Donangelo, C. M.; Farah, A.
Modeling weight loss and chlorogenic acids content in coffee during
roasting. J. Agric. Food Chem. 2010, 58, 1223812243.
(31) Redgwell, R. J.; Trovato, V.; Curti, D.; Fischer, M. Effect of
roasting on degradation and structural features of polysaccharides in
arabica coffee beans. Carbohydr. Res. 2002, 337, 421431.
(32) Oosterveld, A.; Voragen, A. G. J.; Schols, H. A. Effect of roasting
on the carbohydrate composition of Cof fea arabica beans. Carbohydr.
Polym. 2003, 54, 183192.
(33) Robbins, R. J. Phenolic acids in foods: an overview of analytical
methodology. J. Agric. Food Chem. 2003, 51, 28662887.
(34) Nunes, F. M.; Coimbra, M. A. Chemical characterization of
galactomannans and arabinogalactans from two arabica coffee
infusions as affected by the degree of roast. J. Agric. Food Chem.
2002, 50, 14291434.
(35) Nunes, F. M.; Coimbra, M. A. Chemical characterization of the
high molecular weight material extracted with hot water from green
and roasted robusta coffee as affected by the degree of roast. J. Agric.
Food Chem. 2002, 50, 70467052.
(36) Huang, D.; Ou, B.; Prior, R. L. The chemistry behind
antioxidant capacity assays. J. Agric. Food Chem. 2005, 53, 18411856.
(37) Waterhouse, A. L. Determination of total phenolics. In Current
Protocols in Food Analytical Chemistry; Wrolstad, R. E., Ed.; Wiley:
Hoboken, NJ, USA, 2001.
(38) Singleton, V. L.; Orthofer, R.; Lamuela-Raventos, R. M. Analysis
of total phenols and other oxidation substrates and antioxidants by
means of Folin-Ciocalteu reagent. Methods Enzymol. 1999, 299, 152
178.
(39) Homma, S.; Terasawa, N.; Kubo, T.; Yoneyama-Ishii, N.; Aida,
K.; Fujimaki, M. Changes in chemical properties of melanoidin by
oxidation and reduction. Biosci., Biotechnol., Biochem. 1997, 61, 533
535.
(40) Rawel, H. M.; Rohn, S.; Kroll, J. Characterisation of 11S protein
fractions and phenolic compounds from green coffee beans under
special consideration of their interactions a review. Dtsch. Lebensm.
Rundsch. 2005, 101, 148160.
(41) Montavon, P.; Mauron, A. F.; Duruz, E. Changes in green coffee
protein profiles during roasting. J. Agric. Food Chem. 2003, 51, 2335
2343.
(42) Ali, M.; Homann, T.; Rawel, H.; Kreisel, J.; Khalil, M.;
Puhlmann, R.; Kruse, H.-P. Characterization and modeling of the

Real (PEst-OE/QUI/UI0616/2014), and to the QOPNA

Research Unit 62/94 (Project PEst-C/QUI/UI0062/2013;
FCOMP-01-0124-FEDER-037296) by the Foundation for
Science and Technology (FCT) and COMPETE.
Notes

The authors declare no competing nancial interest.

REFERENCES

(1) Ledl, F.; Schleider, E. New aspects of the Maillard reaction in

foods and in the human body. Angew. Chem., Int. Ed. Engl. 1990, 29,
565594.
(2) Hofmann, T.; Bors, W.; Stettmaier, K. Radical-assisted
melanoidin formation during thermal processing of foods as well as
under physiological conditions. J. Agric. Food Chem. 1999, 47, 391
396.
(3) Fogliano, V.; Morales, F. J. Estimation of dietary intake of
melanoidins from coffee and bread. Food Funct. 2011, 2, 117123.
(4) Bekedam, E. K.; Schols, H. A.; van Boekel, M. A.; Smit, G. High
molecular weight melanoidins from coffee brew. J. Agric. Food Chem.
2006, 54, 76587666.
(5) Bekedam, E. K.; De Laat, M. P. F. C.; Schols, H. A.; Van Boekel,
M. A. J. S.; Smit, G. Arabinogalactan proteins are incorporated in
negatively charged coffee brew melanoidins. J. Agric. Food Chem. 2007,
55, 761768.
(6) Nunes, F. M.; Coimbra, M. A. Melanoidins from coffee infusions.
Fractionation, chemical characterization, and effect of the degree of
roast. J. Agric. Food Chem. 2007, 55, 39673977.
(7) Nunes, F. M.; Coimbra, M. A. Role of hydroxycinnamates in
coffee melanoidin formation. Phytochem. Rev. 2010, 9, 171185.
(8) Silvan, J. M.; Morales, F. J.; Saura-Calixto, F. Conceptual study on
Maillardized dietary fiber in coffee. J. Agric. Food Chem. 2010, 58,
1224412249.
(9) Takenaka, M.; Sato, N.; Asakawa, H.; Wen, X.; Murata, M.;
Homma, S. Characterization of a metal-chelating substance in coffee.
Biosci, Biotechnol., Biochem. 2005, 69, 2630.
(10) Gniechwitz, D.; Reichardt, N.; Ralph, J.; Blaut, M.; Steinhart, H.;
Bunzel, M. Isolation and characterisation of a coffee melanoidin
fraction. J. Sci. Food Agric. 2008, 88, 21532160.
(11) Morales, F. J.; Somoza, V.; Fogliano, V. Physiological relevance
of dietary melanoidins. Amino Acids 2012, 42, 10971109.
(12) Moreira, A. S. P.; Nunes, F. M.; Domingues, M. R.; Coimbra, M.
A. Coffee melanoidins: structures, mechanisms of formation and
potential health impacts. Food Funct. 2012, 3, 903915.
(13) Reichardt, N.; Gniechwitz, D.; Steinhart, H.; Bunzel, M.; Blaut,
M. Characterization of high molecular weight coffee fractions and their
fermentation by human intestinal microbiota. Mol. Nutr. Food Res.
2009, 53, 287299.
(14) Bekedam, E. K.; Schols, H. A.; van Boekel, M. A.; Smit, G.
Incorporation of chlorogenic acids in coffee brew melanoidins. J. Agric.
Food Chem. 2008, 56, 20552063.
(15) Perrone, D.; Farah, A.; Donangelo, C. M. Influence of coffee
roasting on the incorporation of phenolic compounds into
melanoidins and their relationship with antioxidant activity of the
brew. J. Agric. Food Chem. 2012, 60, 42654275.
(16) Nunes, F. M.; Cruz, A. C. S.; Coimbra, M. A. Insight into the
mechanism of coffee melanoidin formation using modified in bean
models. J. Agric. Food Chem. 2012, 60, 87108719.
(17) Jilek, M. L.; Bunzel, M. Dehydrotriferulic and dehydrodiferulic
acid proles of cereal and pseudocereal ours. Cereal Chem. 2013, 90,
507514.
(18) Amorim, H. V.; Josephson, V. R. Water soluble protein and
nonprotein components of Brazilian green coffee beans. J. Food Sci.
1975, 40, 11791184.
(19) Hulme, A. C. The isolation of chlorogenic acid from the apple
fruit. Biochem. J. 1953, 53, 337340.
(20) Nunes, F. M.; Coimbra, M. A. Chemical characterization of the
high molecular weight material extracted with hot water from green
and roasted arabica coffee. J. Agric. Food Chem. 2001, 49, 17731782.
7852

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853

Journal of Agricultural and Food Chemistry

Article

interactions between coffee storage proteins and phenolic compounds.

J. Agric. Food Chem. 2012, 60, 1160111608.
(43) Stadler, R. H.; Welti, D. H.; Stampfli, A. A.; Fay, L. B. Thermal
decomposition of caffeic acid in model systems: identification of novel
tetraoxygenated phenylindan isomers and their stability in aqueous
solution. J. Agric. Food Chem. 1996, 44, 898905.
(44) Frank, O.; Blumberg, S.; Kunert, C.; Zehentbauer, G.; Hofmann,
T. Structure determination and sensory analysis of bitter-tasting 4vinylcatechol oligomers and their identification in roasted coffee by
means of LC-MS/MS. J. Agric. Food Chem. 2007, 55, 19451954.
(45) Jaiswal, R.; Matei, M. F.; Golon, A.; Witt, M.; Kuhnert, N.
Understanding the fate of chlorogenic acids in coffee roasting using
mass spectrometry based targeted and non-targeted analytical
strategies. Food Funct. 2012, 3, 976984.
(46) Smrke, S.; Opitz, S. E. W.; Vovk, I.; Yeretzian, C. How does
roasting affect the antioxidants of a coffee brew? Exploring the
antioxidant capacity of coffee via on-line antioxidant assays coupled
with size exclusion chromatography. Food Funct. 2013, 4, 10821092.
(47) Borrelli, R. C.; Visconti, A.; Mennella, C.; Anese, M.; Fogliano,
V. Chemical characterization and antioxidant properties of coffee
melanoidins. J. Agric. Food Chem. 2002, 50, 65276533.
(48) Delgado-Andrade, C.; Rufian-Henares, J. A.; Morales, F. J.
Assessing the antioxidant activity of melanoidins from coffee brews by
different antioxidant methods. J. Agric. Food Chem. 2005, 53, 7832
7836.
(49) Garsetti, M.; Pellegrini, N.; Baggio, C.; Brighenti, F. Antioxidant
activity in human faeces. Br. J. Nutr. 2000, 84, 705710.
(50) Saura-Calixto, F. Antioxidant dietary fiber product: a new
concept and a potential food ingredient. J. Agric. Food Chem. 1998, 46,
43034306.

7853

dx.doi.org/10.1021/jf501510d | J. Agric. Food Chem. 2014, 62, 78437853